Contents
AI in Shopping: Can Chatbots Really Help You Snag the Best Deals?
Artificial intelligence (AI) is rapidly transforming the retail landscape, making its way into everyday shopping experiences. Major sales events, such as Black Friday and Cyber Monday, naturally become significant testing grounds for its practical application. This article delves into whether prominent AI chatbots like ChatGPT, Gemini, and Perplexity can effectively uncover genuine promotions, accurately compare prices, and highlight truly valuable deals for consumers.
The Growing Acceptance of AI for Shopping
Consumer interest in leveraging AI for shopping assistance is on the rise globally. Recent surveys indicate a growing willingness among shoppers to use AI for tasks such as comparing products, prices, and reviews. For instance, a significant percentage of consumers (e.g., over 60% in some regions) express interest in using AI for product and price comparisons, while over half are open to AI-powered shopping assistants. However, there remains a clear boundary: most consumers are hesitant to grant AI direct control over their finances, with only a small fraction (e.g., less than 10%) comfortable with AI completing purchases even with prior approval. This suggests a desire for AI to act as a powerful research tool rather than an autonomous purchasing agent.
Further research highlights the increasing adoption of AI in shopping, particularly during peak sales seasons. Studies show that a substantial majority of shoppers (e.g., nearly 80%) consider making purchases during these periods and are open to using AI support. A considerable portion (e.g., over 30%) already use AI tools for shopping, with another similar percentage planning to do so for upcoming deal hunts.
Navigating Sales: Is That Deal Truly an Opportunity?
The trend of employing AI for pre-purchase research intensifies as major sales events like Black Friday and Cyber Monday approach. This surge is largely driven by what’s known as “promotional fatigue.” Retailers often extend sales over several weeks, forcing consumers to invest more time and effort to verify the authenticity of discounts. Shoppers are wary of being misled or overpaying.
During the Black Friday frenzy, sellers frequently create the illusion of significant savings. A common tactic involves artificially inflating prices several weeks before the sale, then presenting them as “discounted.” This makes the markdown appear more substantial, even if the final price is higher than it was months prior to the sale.
Other deceptive practices include:
- Inflating reference prices: Displaying a supposedly higher original price without any actual prior sales at that price.
- Fake scarcity: Using “last few items” or countdown timers to create a sense of urgency and pressure, compelling consumers to make rushed decisions.
- Dark patterns: These are deceptive user interface designs that trick users into doing things they might not otherwise do, such as automatically adding extra products to a cart or making it difficult to cancel subscriptions.
How We Tested AI in Shopping
To assess the practical utility of AI in shopping, an experiment was conducted using three popular free-version chatbots: ChatGPT, Gemini, and Perplexity. Each chatbot received identical prompts simultaneously to find the best deals on specific products. The objective was to determine if any single AI tool emerged as the most reliable shopping assistant.
The effectiveness of AI tools heavily relies on the quality of the prompt. The initial test began with a relatively short and general command:
“Find the Roborock Qrevo 5AE at the lowest price.”
The choice of this specific product for the search was strategic; it was known that a particular retailer had just launched a promotion for this model. In this initial round, ChatGPT emerged as the clear winner. It was the only AI to identify that the robot vacuum was currently on promotion, while the other chatbots provided significantly higher prices for the device.
However, such a brief prompt doesn’t always yield satisfactory results. To refine the search, a more detailed prompt was used:
“Find the current price of the Samsung Galaxy S26 FE 8/256 GB in [Country/Region, e.g., your local market], list the three cheapest reliable stores, provide the price including delivery, the date of price update, and the source. Do not suggest offers whose availability you cannot confirm.”
For this more complex query, the differences between the chatbots were less pronounced. All three successfully identified recent offers for the smartphone. ChatGPT was particularly notable for its inclusion of visuals and a price monitoring feature. Perplexity, however, pinpointed the absolute lowest price (which was verified to be active) but then struggled significantly with the subsequent task. This experiment highlights the varying capabilities of different AI models, a topic further explored in our analysis of AI integration in automotive systems like CarPlay and Android Auto with Gemini.
The next prompt was even more elaborate, focusing on a coffee machine:
“Find a current and genuine promotion for the NIVONA CafeRomatica 756 in [Country/Region]. Provide a maximum of 5 reliable offers, including: store name, price, delivery cost, total cost, availability, return conditions, a direct link, and the date of update. Assess if the price is favorable compared to the market and historical prices. Clearly mark any out-of-stock or unconfirmed offers. Do not fabricate data. If you cannot confirm something, state it explicitly.”
In this final test, both ChatGPT and Gemini successfully identified a promotional offer for the coffee machine, where applying a specific code resulted in a significant discount (e.g., several hundred currency units). Perplexity, however, failed to include this crucial detail and even struggled to locate the product at a major electronics retailer. The performance differences observed reinforce the continuous evolution of AI models, as seen with advancements like Google Gemini 3.1 Flash’s live AI conversation and voice search capabilities.
Summary: The Future of AI in Shopping
Is a single chatbot sufficient for finding the best deals? Absolutely not. The experiments clearly demonstrate significant variations in results across different AI tools. The quality of outcomes can certainly be improved by crafting more specific and detailed prompts, thoroughly describing the desired product, and narrowing or broadening the search sources as needed.
It’s crucial to remember that the AI’s response is heavily influenced by the underlying language model being used. While free versions often don’t offer model selection, premium subscriptions typically provide access to a wider array of advanced AI models, such as Claude Sonnet, Gemini Pro, or various GPT iterations, each with unique strengths.
Furthermore, AI’s presence extends far beyond chatbots. Artificial intelligence is deeply integrated into:
- Logistics applications
- Price comparison websites
- E-commerce marketplace platforms
- Visual search engines
AI powers various features designed to enhance the shopping experience. For example, it assists product discovery on major online retailers, drives intelligent shopping assistants within delivery service applications, and underpins AI-powered search functions on price comparison sites. Beyond these, there are ongoing trials in agent-based commerce involving prominent retailers and payment providers, aimed at streamlining and automating the purchasing process.
Regardless of the specific AI model or tool, it’s advisable to view artificial intelligence as a powerful assistant for preliminary research rather than a definitive, automated price comparison engine. Placing excessive trust in a single tool might lead to missing out on optimal savings. AI is best used as a starting point, requiring human oversight and verification for the best results.
Frequently Asked Questions (FAQ)
Our tests show that while AI chatbots can identify deals, their effectiveness varies significantly. Some may accurately find promotional prices, while others might miss them or provide outdated information. They are best used as a starting point for research rather than a definitive source.
Retailers often engage in “promotional fatigue” tactics like inflating prices before a sale, using fake scarcity (“last few items”), or employing “dark patterns” (e.g., hidden costs, difficult cancellations). While AI can help sift through numerous offers, human vigilance is still crucial to verify historical pricing and avoid these traps.
To optimize your AI shopping experience, use highly detailed and specific prompts. Include product specifications, desired price ranges, reliable store preferences, and requests for delivery costs, update dates, and sources. Cross-referencing information from multiple AI tools and human verification remains key.
AI is integrated into many facets of shopping, including logistics applications, price comparison websites, e-commerce marketplace platforms, and visual search engines. It powers personalized recommendations, fraud detection, and enhances customer service through virtual assistants.
Source: Internal research, Inside-digital, RP, PRNews
Opening photo: Graphic generated by Google Gemini